As mentioned in the Gearset app fundamentals module, Gearset was built to be smart - it doesn't just compare your metadata, it actually understands it. One key part of this is its problem analyzers. But what is a problem analyzer?

Metadata is complex. Most components do not exist in isolation - they reference each other to create the complex functions that you build in the Salesforce platform. 

For example, when you create a new custom field on an object, you also create a set of permissions defining which users can access that field. Those permissions exist as separate metadata components in the org. As the permissions refer to the field, you can't deploy them without it. This linkage is a called a dependency.

One of the most common challenges of deploying configuration changes in Salesforce is ensuring you've included all of the metadata dependencies. A single missing component can cause the entire deployment to fail. Finding and fixing these dependency issues can be time-consuming and requires a deep understanding of the metadata.

To help you work around these issues, Gearset has a series of algorithms that scan through your metadata before it gets uploaded to Salesforce, and look for any missing dependencies, as well as other potential issues that could result in a Salesforce deployment failure. We call these our problem analyzers, and Gearset has well over 50 of them helping fix different deployment blockers.

The problem analyzers are built with a single purpose: make your deployments more likely to succeed.

Each problem analyzer looks for a specific issue (e.g. a type of missing dependency) and helps you fix it. For example, the master-detail problem analyzer looks for any master-detail relationships, and automatically includes the master field even if it is identical between both environments. This works around a known bug in the Salesforce metadata API and allows our users to successfully deploy these changes where, if using Ant or other tools, they would fail.

The problem analyzers run before any deployment in Gearset, whether done manually or as part of an automated job. Each problem analyzer will run through the selected metadata items and flag up any potential issues you need to address in order to make your deployment succeed.

That's the theory; let's see how they operate in practice. Let's add two fields to your source org, one of which depends on the other.

The first field tracks the discounted value for a product that a member of the sales team might offer to a customer.

Time to create the second, dependent field, to track how much this product discount works out as a percentage.

You've now got two new fields in your source org. One provides an option to discount a product value, the other calculates that discount as a percentage. Time to see how that's represented in Gearset.

Gearset's problem analyzers provide suggestions during the final stages of preparing your deployment. But they're also working behind the scenes all the time by surfacing metadata dependencies right in the comparison viewer.

This dependency tree is a really powerful way to explore and understand dependencies between your metadata, and it's all powered by Gearset's problem analyzers. If you can expand out an item and see dependencies listed for it, that means there's a problem analyzer working behind the scenes to understand that metadata item.

To show how the problem analyzers work during a deployment, let's pretend you didn't know about this dependency between these two fields and just wanted to deploy the percentage field to your target org for testing.

You've just triggered your first problem analyzer! Let's look at what's happening:

For now, proceed without accepting the suggestion.

This is a very simple scenario - in reality you would likely have known about this dependency and included it in your deployment the first time around. But it demonstrates the concept of how Gearset's problem analyzers work. Imagine if you were deploying hundreds of items, some of which were created by colleagues. It would be extremely easy to miss a few crucial items and block a planned deployment. The error messages from Salesforce are not always quite so helpful in tracking down the issue!

Whenever you deploy changes, Gearset will be there providing a helping hand. It'll suggest things to add into your deployment, things you may need to remove, and offer suggestions for things you may need to make directly in your orgs. You don't always have to accept the changes, but in the majority of cases doing so will make your deployments successful.

Problem analysis happens during any deployment in Gearset - whether manual or automated. When running a manual deployment, you can review the suggestions from problem analyzers before continuing. In continuous integration (CI) jobs you can choose to Apply all fixes, choose your pre-built template or to not apply any fixes.
​

If you looked closely at the summary of items to deploy table in the worked example above, you may have noticed that there was something missing - although the DiscountRate field was added by the problem analyzer, the associated Permissions for... item was not. This means that when the field was deployed, it would have had the standard permissions applied to it. 

So why didn't Gearset include the permissions as well?

When looking at what to include in a deployment, there are two broad categories:

Number 1 comes down to a binary result: the deployment succeeds, or it fails. You can't deploy a field without its dependencies. Salesforce will reject it.

Number 2 requires intent and meaning. You can deploy a field without setting the permissions. Would you want to? That's another question! Perhaps you have some unfinished work on the permissions that don't want to deploy yet? Perhaps you're testing what the default settings are in different environments? Perhaps you must include the permissions to avoid exposing sensitive data to the wrong people? There is no way for Gearset to automatically know this and do the right thing.

As a result, Gearset's problem analyzers target number 1. Gearset will help include what you need to make your deployments succeed, you just need to check you've got all the things you want.